import io
import os
import subprocess
import pandas as pd


def call_lastools(tool_name, input, output=None, args=None, verbose=True):
    """Send commands to the lastools library.

    Requires lastools in system path.

    Args:
        tool_name:  name of lastools binary
        input:      bytes from stdout, or path to main input data
        output:     '-stdout' to pipe output, or path to main output data
        args:       list of additional arguments, formatted for lastools
        verbose:    show all warnings and messages from lastools (boolean)

    Returns:
        bytes of output las, if output='-stdout'
        None, if output='path/to/las/file'

    Examples:

    # Convert xyz file to las and pipe stdout to a python bytes object
    las_data = call_lastools('txt2las', input='points.xyz', output='-stdout',
                             args=['-parse', 'sxyz'])

    # Clip las_data with a shapefile, and save to a new las file
    call_lastools('lasclip', input=las_data, output='points.las',
                  args=['-poly', 'boundary.shp'])
    """

    # Start building command string
    cmd = [tool_name]

    # Parse input
    if type(input) == bytes:
        # Pipe input las bytes to stdin
        cmd += ['-stdin']
        stdin = input
    else:
        # Load las from file path
        cmd += ['-i', input]
        stdin = None

    # Parse output
    if output == '-stdout':
        # Pipe output las to stdout
        cmd += ['-stdout']
    elif output:
        # Save output las to file
        cmd += ['-o', output]

    # Append additional lastools arguments, if provided
    if args:
        cmd += [str(a) for a in args]

    process = subprocess.Popen(
        cmd,
        stdout=subprocess.PIPE,
        stderr=subprocess.PIPE,
        stdin=subprocess.PIPE)
    stdout, stderr = process.communicate(input=stdin)

    # Handle errors, if detected
    if process.returncode != 0:
        print("Error: {} failed on {}".format(tool_name,
                                              os.path.basename(input)))
        print(stderr.decode())

    elif verbose:
        # Print addional messages if verbose mode is being used
        print(stderr.decode())

    # Output piped stdout if required
    if output == '-stdout':
        return stdout
    else:
        return None


def extract_pts(las_in, cp_in, survey_date, beach, args=None, verbose=True):
    """Extract elevations from a las surface based on x and y coordinates.

    Requires lastools in system path.

    Args:
        las_in:       bytes from stdout, or path to main input data
        cp_in:        point coordinates with columns: id, x, y, z (csv)
        survey_date:  survey date string, e.g. '19700101'
        beach:        beach name
        args:         list of additional arguments, formatted for lastools
        verbose:      show all warnings and messages from lastools (boolean)

    Returns:
        Dataframe containing input coordinates with extracted elevations

    Examples:

    # Extract elevations from 'points.las', using control points from 'cp.csv'
    # Specify control point format as: id, x, y, z ('-parse', 'sxyz')
    # Only use points classified as 'ground' ('-keep_class', '2')
    extract_pts('points.las', 'cp.csv', survey_date='20001231', beach='manly',
                args=['-parse', 'sxyz', '-keep_class', '2'])
    """

    # Assemble lastools arguments
    if args:
        args = ['-cp', cp_in] + args
    else:
        args = ['-cp', cp_in]

    las_data = call_lastools(
        'lascontrol', input=las_in, output='-stdout', args=args, verbose=False)

    # Load result into pandas dataframe
    df = pd.read_csv(io.BytesIO(las_data))
    
    # Create empty dataframe if no control points intersect point cloud
    if (df.iloc[:, 0] == '-').all():
        df = pd.read_csv(cp_in)
        df['diff'] = '-'
        df['lidar_z'] = '-'

    # Tidy up dataframe
    df = df.drop(columns=['diff'])
    df['lidar_z'] = pd.to_numeric(df['lidar_z'], errors='coerce')
    df['Beach'] = beach
    df = df[[
        'Beach', 'ProfileNum', 'Easting', 'Northing', 'Chainage', 'lidar_z'
    ]]

    # Rename columns
    new_names = {
        'ProfileNum': 'Profile',
        'lidar_z': 'Elevation_{}'.format(survey_date),
    }
    df = df.rename(columns=new_names)

    return df


def update_survey_output(df, output_dir):
    """Update survey profile output csv files with current survey.

    Args:
        df:          dataframe containing current survey elevations
        output_dir:  directory where csv files are saved

    Returns:
        True if current survey is latest, otherwise False
    """
    # Merge current survey with existing data
    profiles = df['Profile'].unique()
    for profile in profiles:
        csv_name = os.path.join(output_dir, profile + '.csv')

        # Extract survey data for current profile
        current_profile = df[df['Profile'] == profile]
        try:
            # Load existing results
            master = pd.read_csv(csv_name)
        except FileNotFoundError:
            master = current_profile.copy()

        # Add (or update) current survey
        current_survey_col = current_profile.columns[-1]
        master[current_survey_col] = current_profile[current_survey_col].values

        # Prepare output directory
        try:
            os.makedirs(output_dir)
        except FileExistsError:
            pass

        # Ensure survey dates are in correct order
        elev_cols = sorted([col for col in master.columns if 'Elevation' in col])
        other_cols = [col for col in master.columns if 'Elevation' not in col]
        master = master[other_cols + elev_cols]

        # Export updated results
        master.to_csv(csv_name, index=False, float_format='%0.3f')